Extraction of lane markings using orientation and vanishing point constraints in structured road scenes

Abstract

Extraction of lane markings is still a challenge due to the poor quality of lane markings and interferences from external circumstances. In this paper, we utilize line segments as low-level features to detect lane markings on structured road scenes. Our novel algorithm can be highlighted in four items as follows. Firstly, a road surface region is reasonably located using adaptive segmentation method, and then most of the interferences from the external circumstances are eliminated by Laplacian filter. Secondly, the line segments detected by the line segment detector are applied to represent the structural information of lanes. Thirdly, non-lane candidate markings are removed through orientation and vanishing point constraints. Finally, the lanes are accurately extracted from the remaining candidate lane marking. Experimental results on complex structured road scenarios in urban streets are shown and the effectiveness and robustness of our novel algorithm are underpinned by the experimental results.